Philippine Institute for Development Studies Surian sa mga Pag-aaral Pangkaunlaran ng Pilipinas
Children and the Labor Force Participation and Earnings of Parents in the Philippines Aniceto C. Orbeta Jr. DISCUSSION PAPER SERIES NO. 2005-20
The PIDS Discussion Paper Series constitutes studies that are preliminary and subject to further revisions. They are being circulated in a limited number of copies only for purposes of soliciting comments and suggestions for further refinements. The studies under the Series are unedited and unreviewed. The views and opinions expressed are those of the author(s) and do not necessarily reflect those of the Institute. Not for quotation without permission from the author(s) and the Institute.
September 2005 For comments, suggestions or further inquiries please contact: The Research Information Staff, Philippine Institute for Development Studies 3rd Floor, NEDA sa Makati Building, 106 Amorsolo Street, Legaspi Village, Makati City, Philippines Tel Nos: 8924059 and 8935705; Fax No: 8939589; E-mail: publications@pidsnet.pids.gov.ph Or visit our website at http://www.pids.gov.ph
Children and the Labor Force Participation and Earnings of Parents in the Philippines
Aniceto C. Orbeta, Jr. June 2005
Abstract This paper demonstrates how family size can be an important contributor to poverty in the Philippines. It examines one of the mechanisms behind this link by focusing on the relation between number of children and the decision to seek a job and parents' wage earnings. It surveys the international literature to establish how the problem has been approached and what the results are for other countries. It then formulates and tests a model using a nationally representative household survey data for the Philippines to explain what determines the decision to seek a job and the earnings of both mothers and fathers. The model specifically considered the endogeneity of the number of children in both the labor force participation and the earnings equations. Keywords: Family Size, Labor Force Participation, Earnings, Philippines
Children and the Labor Force Participation and Earnings of Parents in the Philippines1 Aniceto C. Orbeta, Jr.2 June 2005 1. Introduction How children affects the labor force participation and earnings of mothers and fathers can spell the difference on whether additional children are can expect the needed care or not. When parents exert more effort with additional children, then their impact of on the welfare of the family will be mitigated. When the opposite happens, not only will additional children not get the needed support, they will also cause the deterioration of welfare of the other members of the household as resources are spread to more members. It is therefore important to quantify the impact of children on the work effort and earnings of their parents. Even though the average education of women in the Philippines is higher, their labor force participation is significantly lower than her Asian neighbors. One explanation that can be put forward is, of course, the inconsistent growth rate the country has been experiencing for a couple of decades now. Another, perhaps commonly forgotten reason, is that while her neighbors have successfully brought down their fertility rates, the Philippines has failed to reduced its fertility rate as fast as say Thailand, Indonesia and Vietnam. The burden of many children can limit the ability of mothers to avail of work opportunities thus stalling the rise in the work uptake of Filipino women. This paper formulates and estimates a model of the determinants of the labor force participation and earnings of mothers and fathers with the number of children as one the explanatory variables that include individual, household and community characteristics. It uses the nationally representative 2002 Annual Poverty Indicators Survey in the analysis. This is one of the few papers that recognized and thoroughly tested the endogeneity of the children in these equations. This, however, did not produce positive results with the dataset used. This result lends support to the validity of using estimates that consider the number of children exogenous in the data set used for the study. The estimation results generated rich results that provided quantitative estimates of the impact of children on the labor force participation and earnings of parents. The estimates point to highly regressive impact of additional children on Philippine households. The paper is divided as follows. The next section provided a selective review the previous literature. A presentation of the methodology, instrument and data is provided
1
This paper also appeared as ADBI Discussion Paper No. 30. Senior Research Fellow, Philippine Institute for Development Studies (aorbeta@mail.pids.gov.ph). This paper was written while the author was a Visiting Researcher at the Asian Development Bank Institute, Tokyo. Opinions expressed here are solely of the author does not necessarily reflects the view or policies of the Asian Development Bank Institute nor of the Philippine Institute for Development Studies. This paper has benefited from the comments of John Weiss, Haider Khan, and Peter McCawley. Research assistance of Janet Cuenca, Keiko Sasaki, Mihoko Saito, Reiko Nishiura and Nami Sampei are gratefully acknowledged. All errors, however, are solely the responsibility of the author. 2
2 next. The estimation results are presented in section four. The final section provides a summary and implications for policy. 2. Review of Previous Studies Browning (1992) provides a full section in his paper reviewing the US literature on the impact of children on female labor supply up to the early 1990s. The observations he made on the literature was that it is clear that there is a strong negative correlation between the presence of young children and the labor supply of their mothers measured either as labor force participation or labor hours. The relationship, however, is not as clear as one moves from correlation to causal relationships. This is because the relationship between work and child bearing is very complex conceptually and also to estimate empirically, as paper has shown. One important point highlighted in the Browing (1992) is that labor supply equations that does not include the children variables as regressors are by implication estimating a reduced form. To illustrate the importance of recognizing the endogeneity of the fertility the review indicated that those that consider fertility as endogenous have yielded not significant or even positive relationships putting to question results that show negative relationship. He expressed frustration that the studies that considered the endogeneity of fertility did not also show OLS results to be able to make a comparison. The results in Rosenzweig and Wolpin (1980) show that not instrumenting for the fertility variables underestimates the negative impact of exogenous changes in fertility. The results in Angrist and Evans (1998), on the other hand, show that controlling for the endogeneity either by the use of sex of first two births or twins as instrument yields negative impact of the number of children on the labor supply of mothers but the impact is much more subdued than those obtained from OLS estimates or the opposite of what Rosenzweig and Wolpin (1980) obtained. Gangadharan and Rosenbloom (1996) adds that the negative impact on hours increase as the magnitude of mothers enter the labor market and that while the negative impact on earnings in 1980 is temporary, the 1990 results show continued depressed earnings even after the labor supply effects had disappeared. Thus, while the negative effect of children in mother’s labor supply in these three studies, the direction of the impact of recognizing endogeneity of fertility variables in the labor supply equation is not clearly established. In the case of the father’s labor supply, Angrist and Evans (1998) did not find significant effect of children. Lundberg and Rose (2002), however, found positive effect of children. In addition, they found that the effect of children on male labor supply is substantially larger when endogeneity is taken into account lending support to the Rosenzweig and Wolpin (1980) result. Vere (2005) adds the new 2000 Census data to the 1980 and 1990 used in Angrist and Evans (1998). He finds that men respond to additional children by increasing earnings lending support to Lundberg and Rose (2002). It is also noted that specialization in husbands’ and wives’ roles in raising children has become more pronounced over time. In the case of the husband, not only is the sign of the impact of children is not clear but also the direction of the effect of controlling for the endogeneity of the number of children. If the relationships between children and the parents labor supply and earnings are not clear in developed county literature, the same is true in less developed countries with the little evidence that is available.
3 The only study, to the knowledge of the author, using Philippine data that considers the endogeneity of children in women’s work hours and earnings is Adair, et al. (2002). Using the Cebu Longitudinal Health and Nutrition Survey from 1983-1991, the paper models the joint decision of sector of work choice (wage, piece or self-employed) and earnings (or hours of work) with children included in the later equation. The study finds that children negatively affect hours and earnings in what the authors deemed as a form of a “child tax”. They find that for the change in earnings equation the number of additional live births has significant effect but children less than 2 years is not. For the change in labor hours equation, it is the number of children less than 2 years that is significant. Child bearing, however, was not found to affect sector of employment. The review of other studies using Philippine data that assumes fertility variables as exogenous show mixed results. Using survey data from Laguna province, Quizon-King (1978) found that while the number of children does not significantly affect the market time of both mother and father, the presence of children 6 years and below increases the home production time of mothers. Garcia (1990) also found that the percentage of children below six years old significantly increases the time for home tasks and decreases market work for wives while the opposite is true for husbands. Popkin (1983) found that the number of children below 6 yeas old decreases the leisure time of rural mothers; the presence of children one year old and below decreases their market production time while the presence of children 1-6 years old increase their home production time. King and Evenson (1983) found that the presence of children six years old and below increases the home time of wives. For husbands, the presence of children one year and below positively affects his home time. When the age structure of the children was controlled the effects turned out to be small. Market time, however, of both husband and wife, is not affected by the presence of children. Finally, using cross tabulation analysis from the Cebu data Tiefenthaler (1997) has confirmed earlier results on the negative relationship between child bearing and labor supply of mothers. However, it was pointed out that that for mothers with previous children, labor market hours 14 months postnatal is the same prior to the sample birth although this is lower for first-time mothers. For fathers, birth does not significantly affect their labor market hours. In addition, even if mothers decrease labor supply within the year after birth, fathers do not increase their market time to compensate for lost income. Birth only slightly increase father’s time in childcare and only if there are not other children. It was also found that births increase the childcare and market time of older daughters (13-17 years). Thus, it appears that older daughters, rather than fathers, substitute for the lost market time of mothers due to birth. Evidence from other developing countries also had varied results. Using the 1992 DHS for Morocco Assaad and Zoari (2003) found that in urban Morocco the presence of school-age children significantly reduces participation of women in all types of wage work. This particular study uses a three-stage sequential modeling of marriage, child bearing and labor force participation with predicted values of the dependent variable used in the subsequent stage to control for the endogeneity. Multivariate analysis using data in Urban Pakistan and assuming fertility variables as exogenous showed differential effects by sex of children (Cochrane, Kozel and Alderman, 1990). Males 7-14 and over 14 reduce the participation of women while females 7-14 and over 14 significantly increase the labor force participation of women. The magnitude of the effects, however, was small. The authors added that these implies that males provided alternative source while females free up women’s home time. In the case of rural Thailand Poshisita et al (1990), using cross tabulation analyses supported by results
4 from focused grouped discussion, finds that while children does not prevent rural Thai women from working, they cause work interruption and interfere with there their economic activity. 3. Methodology, the Instrument and Data 3.1 Methodology To determine the impact of the number of children in the household on the labor supply and earnings of parents we estimate these relationships by recognizing the endogeneity of the number of children. The importance of recognizing the endogeneity of children in the labor supply and earnings equation of parents has been highlighted in the previous section. We follow Angrist and Evans (1998) in assuming balanced sex-mix and using same sex of the first two-births as the instrument. The validity of this instrument for the number of children is explained after a discussion of the empirical specification. While it would have been desirable to include labor hours, the data that we use does not have this information. For labor supply, therefore, we only estimate an equation for the labor force participation of parents. Labor Force Participation of Parents. The labor force participation rate equation we estimate is the following model:
(1) l = α 0 + α1 w1 + a 2Y + α 3n + Xα 4 + ε ( 2) n = β 0 + β1 z + Xβ 2 + µ Equation (1) is a typical labor supply equation where l is the labor force participation of the parent, w is wage, Y is other (non-wage) income received by the household, n is the number of children, and X as the set of control variables and e the disturbance term. The vector X would usually include age and education. The estimation methodology is as follows. Since (1) is a dichotomous choice model, we will use probit to estimate the model. But the endogeneity of n will result in a biased estimate. We therefore test for the endogeneity of n using the suggestions in Rivers and Vuong (1988). They proposed a two-stage probit where the estimated error from the first-stage regression is added as an explanatory variable in the second-stage probit regression to obtain a consistent estimate. The pointed out that the coefficient of the error term will constitute a test for the endogeneity of n. Rivers and Vuong (1988) indicated that adjustment is needed for the variance-covariance matrix in the second stage probit to get asymptotically correct errors. Bollen, Guilkey and Mroz (1995) however, established through Monte Carlo simulations that results are not readily affected by using the asymptotically correct standard errors3. Another method is the efficient4 full-information maximum likelihood (FIML) estimation of the two equations and 3
Bollen, Guilkey and Mroz (1995) estimates a very similar problem of the endogeneity of the desired additional children on the contraceptive use equation. 4 Although FIML produces efficient estimates, there are a couple of limitations inherent in the FIML. One, obviously it is much more difficult to estimate although this is increasingly not too much of a concern as most statistical package can be programmed to generate FIML estimates. In fact there is an available
5 directly testing the significance of the correlation between the error terms in the number of children equation and the labor force participation equation. Both of these methods are used to establish the endogeneity of the number of children variable in the labor force participation equation. If n is found to be endogenous we will use the estimation that will give us more precise (higher significance) estimate for the variable of interest – the number of children, otherwise, we use the simple probit results. Earnings of Parents. Similarly, to determine the impact of the number of children on their parents’ wage earnings we estimate an augmented Mincerian equation of the following form
(3) ln w = α o + α1age + α 2 age 2 + α 3educ + α 4 n + Xα + ε ( 4 ) n = β o + β 1 z + Xβ 2 + µ The estimation used the standard Mincerian equation for estimating the earnings of parents with the number of children and location dummies added. This is essentially adopted from Angrist and Evans (1998). We instrument for n using the sex of the firsttwo births as explained in the next section. In addition, w may not be observed for those who did not work requiring adjustment for the censoring. The estimation methodology is similar to the one described above except that the second stage regression in this model is censored. Since the earnings will be zero both for the non-wage workers and the non-workers, Tobit estimation is used to take account for the censoring in the earnings equation. We test for the endogeneity of n in the earnings equation. Smith and Blundell (1986) suggested a two-stage Tobit to determine the endogeneity of the variable in the structure described above. Specifically, the estimated error term from the first-stage regression is added as a variable in the earnings equation to arrive at a consistent estimate. Thi A significant coefficient for the estimated error term implies endogeneity of n. If the n is found to be endogenous, we use the results of the two-stage Tobit estimation. Otherwise, we use the simple Tobit estimation results. 3.2 Balanced Sex-Mix as an Instrument There are not too many instruments that one can find for the number children in household models. Most of the likely candidates such the household income, education of the parents or age of marriage are also related to the dependent variable of interest such as labor force participation of parents, savings or education of children, rendering these inappropriate as instruments. Recent research using US data such as Angrist and Evans (1998) has used the hypothesis that families prefer to have balanced sex-mix of children as an instrument for the number of children. The Philippines is one of the countries in Asia where a balance sex-mix are found to have prevailed in contrast to countries in South and Eastern Asia where indications for son preference is often found (Wongboonsin and Ruffolo, 1995). Early literature that confirms preference for balanced sex-mix in the Philippines is found in Stinner and Mader (1975). The other instruments routine in Stata called probitiv that implement FIML estimation of the model above described in Filmer and Lokshin (nd). The second is the natural consequence of system estimation; that is any bias resulting from misspecification in the one of the equation is transferred to the whole system. Limited information estimates, such as the two-stage probit, limits the bias to the mis-specified equation only.
6 that are available are limited by their applicability only in very specific circumstances. The occurrence of twins have been also been used as instruments again using US data first in Rosenzweig and Wolpin (1980a) and in subsequent studies such as Angrist and Evans (1998). A much more recent application was done for the US (Vere 2005), for Romania (Glick, Marini and Sahn, 2005) and for Norway (Black et al, 2004). Sonpreference in Korea was also used as an instrument for the fertility for instance in Lee (2004). Finally, another instrument would be an exogenous policy change that could affect child bearing. Quian (2004), for instance, used the relaxation of the one-child policy in China that allows rural households to have another child if the first child is a girl. Viitanen (2003), on the other hand, used the large-scale giving out of vouchers for privately provided childcare in Finland. In the case of the balanced sex-mix hypothesis, the fact that families do not have control over the sex of their children makes same sex for the first two children virtually a random assignment. As argued in Angrist and Evans (1998) using same sex as an instrument will allow a causal interpretation. It should be noted, however, that the downside of this instrument is that it will render families that has less than two children unusable for analysis. While this maybe a serious problem in low fertility areas, this may not be in the case of the Philippines where the average number of children exceeds four. To check on the validity of this instrument, Table 7 provides a cross tabulation of the average proportion of families that have additional children and the average number of number of children by sex of their first two children for 24,000 families that have two or more children using the APIS 2002 dataset. The table shows that 67.4% families that had one male and one female for their first two children had another child while 71.8% had another child when the have same sex for their first two children or a difference of more than 4%. In terms of average number of children, this is 3.49 as against 3.61 or an average difference of a little over 0.12 children. These average differences are statistically significant under conventional level of significance. Comparing this with Table 3 and 5 in Angrist and Evans (1998) one can observe several differences. The difference in the proportion of families having a third child for the two groups of families is smaller and the standard error is larger. In the case of the difference in the average number of children, the difference is larger but so is the standard error. This is not unexpected given the larger family size in the Philippines and the expected larger dispersion of the distribution. Consequently, the implied t statistics in Table 7 are not as large as those in Angrist and Evans (1998) indicating that discrimination generated from the same-sex instrument may not be as strong as those obtained using US data. 3.3 Data Sources The data on individual and household characteristics and location characteristics were taken from the 2002 Annual Poverty Indicator Survey (APIS). The APIS is a rider survey to the July round of the quarterly Labor Force Survey (LFS) conducted by the National Statistics Office (NSO). The 2002 round is the third of the APIS series conducted by the NSO. The other two were conducted in 1998 and 1999. It provides basic demographic information on all members of the household as well as household amenities. Income and expenditure for the past 6 month period preceding the survey are also gathered. All monetary values such as wage and non-wage income are deflated using provincial consumer price indices compiled by the Price Division of the NSO. This is done to control for inter-provincial price variability.
7
The unemployment rate is computed as the domain level average unemployment rate using APIS data. 3.4 Descriptive Statistics Table 2 shows the proportion of mothers and fathers working by per capita income quintile and by number of children. The proportion of mothers working shows that more mothers in higher income households work both for all types of work and for paid work. The opposite appears to be true for fathers. One way to explain the difference is that richer households are able to pay for house helps freeing them to participate in the labor market and still contribute to household income. This may not be the case for mothers from poorer households. There is no clear explanation for the lower labor force participation of fathers from richer household except perhaps that they may be earning more from other sources. In terms of by number of children, as expected, mothers with smaller number of children work more. The same is true for fathers. It is also noteworthy that unpaid work (the difference between all types of work and paid work) is about 20 percent for mothers. Table 3 provides the descriptive statistics of the variables used in the estimation. The average number of children is about 3.5. The average number of years of education is slightly higher for mothers than fathers at 9.2 versus 9.0, respectively. This is not a surprising phenomenon in the case of the Philippines. About 50% of the households have children below usual primary school aged (6 years). 4. Estimation Results 4.1 Labor Force Participation of Mothers Both the endogeneity tests using either the two-stage probit or the FIML all show that the number of children using boys and both girls or same sex5 or in the first two births as instruments did not show significance. Table 4 shows, the error in number of children did turn out to be significant with z-value of 1.4. In addition, the test of the correlation coefficient from the FIML estimation for both the mother and father labor force participation equation also turned out to be insignificant with chi-square value of 0.2 (Table 5). Both of these imply that for this particular data set the endogeneity of the number of children is not established. This also lends support to the validity of using simple probit estimates. Subsequent discussion will then focus on the simple probit estimates. The simple probit estimates of the determinants of the labor force participation of mothers are given in Table 6. The labor force participation of the mothers are expected to decline by 0.926 percentage points with each additional child. This slightly rises to 0.96 percentage points when only unpaid work is excluded. Another noteworthy result relative to the impact of children on the labor force participation of their mothers is the 5
Not shown. This is from the marginal effects column. Note that probit is a non-linear model so the marginal effects, rather than the coefficients, provide the estimate of the impact of the change in the probability of working to a change in the dependent variable. 6
8 impact of the presence of children below normal school age – 0 to 5 years old. Table 5 shows that the presence of young children below normal school age reduces the probability of their mother working by a considerably higher 7.8 percentage points when all types of work are considered and 5.7 percentage points when unpaid work is excluded. This confirms many of the results in previous studies on the negative impact of children on the labor force participation of mothers. The other significant determinant variables of mother’s labor force participation are her age, education, wage income of the father and unemployment rate. The age of the mother was entered as a quadratic to capture non-linear effects. The signs of the coefficients confirm the expectation that that labor force participation of mothers rises at a declining rate with age. Education is a positive determinant as found in many other studies. The higher the wage income of the father the lesser is the likelihood that a mother would be working. Interestingly, the estimates shows that a thousand increase in the father’s wage income per capita7 would have an equivalent depressing effect of an additional child. Higher unemployment rate discourages mothers from looking for work lending support to the discouraged worker hypothesis. Contrary to expectation, nonwage income per capita of the household is a significant positive determinant of the labor force participation of mother for all types of work although it has the expected negative sign in the paid work equation. To determine the differential impact of children on the labor force participation across income classes, the number of children variable was interacted with the per capita income quintile dummy variables. The result of the estimation is given in Table 7. For all types of work, the interaction variables are not significant except for the top two quintiles. For paid work, however, all the interaction terms are significant. Table 8 provides the summary of the impact on the labor force participation of mothers by per capita income quintiles expressed as a percentage of the recorded participation rates. For all types of work, mothers from the bottom three quintiles will reduce the proportion working by an average of over 2 percent for each additional child. For the top tow quintiles, however, the impact is positive implying that more of them will work with an additional child. This could even reach as high as about 7% for the richest quintile. In the case of paid work, the pattern is similar except that the impacts are much larger in magnitude. About a 6% decline for each additional child for mothers in the bottom quintile and more than 8% increase for each child for mothers in riches quintile. This differentiated impacts undoubtedly provides a richer view of the impacts than just the average impacts. 4.2 Labor Force Participation of Fathers Similar to the results in the mothers labor force participation equation, the test results for endogeneity of the number of children in the fathers labor force participation equation also had insignificant results. The coefficient of the estimated error term in the first stage regression did not turn out to be significant in the second stage labor force participation equation with z value of 0.16 (Table 9). Similarly, the test of the correlation coefficient using FIML also turned out to be insignificant with chi-square value of .03 (Table 5). These lend support to the validity of using simple probit results. The results for the labor force participation of fathers show that on the average the number of children does not affect their labor force participation (Table 10). It is negative 7
Or about five thousand for a family of five
9 but not significant. This means that fathers, on the average, don’t try to find work when a child is added in the family. Similar variables that are significant in the mother’s labor force participation equation are also significant in the father’s labor force participation equation. Age of the father has a rising at a declining rate effect. One surprising results is the negative and significant coefficient for the years of education of the father. Perhaps, this may mean that highly educated fathers are earning more from other sources. Non-wage income is not a significant determinant. Like the mothers, discourage worker hypothesis also works for fathers, i.e., with higher unemployment they tend not to look for work all other things equal. Again to determine the differential impact across income classes, the number of children of children variable was interacted with the per capita income quintile dummy variables. The estimation results are also shown in last four columns of Table 10. All the coefficients of the interaction terms are significant. Since the base category is not significant it will considered zero. The summary of the impact expressed in terms of the recorded proportion of fathers working is also provided in Table 8. For the poorest quintile there would be no effect. The impacts for the upper income quintiles are all positive with 0.3, 0.6, 0.4, and 1.2 for the 2nd, 3rd,4th and 5th quintiles, respectively. Thus, there is slight positive effect for the labor force participation of fathers in the upper income quintiles. 4.3 Earnings of Mothers The endogeneity test using two-stage tobit shows that the number of children is not endogenous in the earnings equation for this particular data set as is seen in the labor force equation. The coefficient of the estimated first-stage error term is not significant with t value of 0.63 (Table 11) lending support to the validity of using ordinary tobit estimates. Thus, the subsequent discussion will refer only to the ordinary tobit results. Table 12 shows the results of the Tobit and OLS estimates for the earnings of mothers. The number of children is found to negatively affect the earnings of mothers on the average. But when one looks at the coefficient of the interaction terms with per capita income, the negative impact is only for the bottom two quintiles8. The upper three quintiles have positive impacts and this is roughly consistent with the results for the labor force participation. It is easier to look at the impact as a percentage to recorded incomes and in absolute value.9 These are shown in Table 13. The average effect is about a 5% decline in income or about 1 thousand pesos from the six-month earnings per additional child. For the bottom two quintiles, the impact is about -13% for the poorest and -7% for the lower middle quintile per additional child. These translate to about a reduction of about 700 and 600 pesos to the semesters wage income, respectively. For the top three quintiles, the impacts are positive: 2%, 15% and 33% for the middle, upper middle and richest quintile, respectively. This means an addition of 360, 6,200 and 25,736 pesos to the semesters wage income for the corresponding quintiles, respectively, per additional child. 8
There are three things to note in computing the impact of the exogenous: (a) this is a tobit model so the marginal effects has to consider censoring, the marginal effects columns computes the unconditional values, (b) note that the dependent variable is in natural log so the marginal effects computed are in percentage terms. 9 Since the estimation uses deflated values, these are inflated back to the 2002 values using the price index.
10
All the other variables, age, education, residing in urban areas and regional dummies are significant in determining the wage income of mothers. Wage income rises with age in a decreasing manner. Education positively affects earnings. There is on average higher earning in urban areas. Except for ARMM, which is unexpected, mothers in all other regions earn lower than those in the national capital region. 4.4 Earnings of Fathers Again the endogeneity test using the two-stage tobit did not yield significant results. The coefficient of the estimated first-stage error term is not significant with a t value of 0.83 (Table 14) lending support to the validity of using the ordinary tobit results. This is what we use in subsequent discussions. The result shows that on average, the earnings of fathers are positively affected by the presence of children (Table 15). This is in contrast to the impact on labor force participation that showed no significant affects. Perhaps father are more serious in finding paying or higher paying jobs rather than just any job with additional children. The impact as a percentage of recorded income and in absolute value is also given in Table 13. There is an average increase of about 1% in income of fathers or an addition of about 233 pesos to the semestral wage income. The interaction terms between the number of children and per capita income quintile dummies also all turned out to be significant. Considering all of these, the negative impact remains for the bottom quintile with about a 6% decline in wage income or about a 76 pesos reduction in semestral income per additional child. For the lower middle up to the richest quintile, the effect is positive from 93 to 25,538 pesos addition to the semestral wage income for the lower middle to the richest quintile per additional child. Judging from these numbers, even though the impact is positive, this is hardly enough to pay for the marginal increase in expenditure due to the new child except perhaps for the richest quintile. The other variables have similar performance with those in the mothers’ earnings equation. Earnings rise with age at a declining rate. Education positively affects income. There is a higher average earnings of fathers in urban areas. Earnings of fathers in the national capital region are also higher than those in the other regions without exception. 5. Summary and Policy Implications The paper presented and estimated a model of the determinants of the labor force participation and the earnings of parents with the number of children as one of the determinants among the usual individual, household and community characteristics. The estimation strategy tests first for the endogeneity of the number of children as suggested in the literature using the sex of the first two births as instruments as suggested in Angrist and Evan (1998). Using both two-stage probit and FIML, the tests failed to establish the endogeneity of the number of children in the data set both in the labor force participation equation. Similarly, using two-stage tobit, the endogeneity of the number of children in the earnings equations was not also substantiated. These lend support to the use of the simple probit results in the labor force participation equation and simple tobit in the earning equations. The results show that on average the impact of children is negative on the labor force participation of mothers and insignificant in the labor force participation of fathers. To determine the differential impact across income quintile, the number of children was interacted with the per capita income quintile dummy variables. This generated a richer
11 result. It was shown that the negative impact of children on the labor force participation of mothers is only found in the bottom three quintiles. For the top two quintiles the impact is positive. Perhaps the mothers in the top quintiles are able to pay for child-care, e.g., house helps, so they are free to work even with additional children. In the case of fathers, the impact is insignificant only in the bottom quintile. In the upper four quintiles the impact is positive, i.e., fathers work more with additional children. In relative terms, the impact in the case of the fathers is much more subdued than it is with the mothers which is not surprising given that mothers have the primary responsibility in child rearing. Turning to the impact on earnings, the average impact is again negative of mothers but this time positive for fathers. The negative impact on earnings is only found in the bottom two quintiles while in the top three quintiles the impact is positive. For fathers, even though the average impact is positive, the impact on the bottom quintile is still negative. The positive impact for the upper four quintiles accelerates as one goes up the income quintiles. From the foregoing, it appears that there is a regressive impact of additional children on the labor force participation and earnings of parents. Combining the results on earnings, the bottom quintile has a double negative impact with each additional child --the mother as well as the father has reduced wage income. For the lower middle quintile, there is an offsetting effect although the increase in the father’s income is not enough to cover the loss in the mother’s income. For the upper three quintiles there is a double positive income effect. The results point to important implications for policy. Owing the regressive impact of additional children, government needs to train family planning assistance at the bottom quintile where there is double negative impact of additional children. In the short run there is a need to assist this group in achieving their fertility goals. Advocating for smaller family size maybe necessary in the long run. The design of any employment or livelihood assistance needs to consider the burden of children, in general, and preschool children, in particular, as these are shown in this paper to limit the ability of mothers to take up work.
12 References Adair, L., D. Guilkey, E. Bisgrove and S. Gultiano (2002) “Effect of childbearing on Filipino women's work hours and earnings,” Journal of Population Economics, 15(), 625645. Assaad, R. and S. Zoari (2003) “Estimating the Impact of Marriage and Fertility on Female Labor Force Participation when Decisions are Inter-related: Evidence from Urban Morocco” Topics in Middle Eastern and North African Economics, electronic journal, 5. http://www.luc.edu/publications/academic/. Cochrane, S., V. Kozel and H. Alderman (1990) “Household Consequences of High Fertility in Pakistan,” World Bank Discussion Paper 111. Gangadharan, J. and J. Rosenbloom (1996) “The Effect of Child-Bearing on Married Women’s Labor Supply and Earnings: Using Twin Births as a Natural Experiment,” NBER Working Paper 5647. Garcia, M. (1990). Resource Allocation and Household Welfare: Study of the Impact of Personal Sources of Income and Food Consumption, Nutrition and Health in the Philippines. Thesis submitted to the Institute of Social Studies, The Hague, The Netherlands. Lundberg, S. and E. Rose (2002), “The Effects of Sons and Daughters On Men’s Labor Supply and Wages”, The Review of Economics and Statistics, 84, 251-268. King, E. and R. Evenson (1983). “Time Allocation and Home Production in Philippine Rural Households,” in M. Buvinic, et al. (eds.) Women and Poverty in the Third World, Baltimore: The Johns Hopskins University Press. Mroz, T. (1988) “Sensitivity of an Empirical Model of Married Women’s Work to Economic and Statistical Assumptions,” Econometrica, 54(4), 765-799. Popkin, B. (1980). “Time Allocation of the Mother and Child Nutrition.” Ecology of Food and Nutrition, Vol. 9, pp. 1-14. Poshisita, C. N. Havanon, J. Knodel, and Werasit Sittitrai (1990) “Women’s work and family size in rural Thailand,” Asia-Pacific Population Journal 5, 31-54. Quizon-King, E. (1978). “Time Allocation and Home Production in Rural Philippine Household,” The Philippine Economic Journal, Vol. 17, No. 1&2, pp. 185-202. Tiefenthaler, J. (1997). “Fertility and Family Time Allocation in the Philippines,” Population and Development Review, Vol. 23. Issue 2, 377-397. Rosenzweig, M. and K. Wolpin (1980) “Life-cycle Labor Supply and Fertility: Causal Inferences from Household Models,” Journal of Political Economy, 88(2), 328-348. Vere, J. (2005) “Life Cycle Effects of Fertility on Parents’ Labor Supply,” University of Hong Kong.
13 Angrist, J. and W. Evans (1998) “Children and their Parents’ Labor Supply: Evidence from Exogenous Variation in Family Size,” American Economic Review 88(3), 450-477. Black, S. , P. Deveruex and K. Salvanes (2004) “The More the Merrier? The Effect of Family Composition on Children’s Education,” NBER Working Paper No. 10720. Bollen, K. D. Guilkey and T. Mroz (1995) “Binary Outcomes and Endogenous Explanatory Variables: Tests and Solutions with an Application to the Demand for Contraceptive Use in Tunisia,” Demography, 32(1), 111-131. Browning, M (1992) “Children and Household Economic Behavior,” Journal of Economic Literature 30(3), 1434-1475. Filmer, D. and M. Lokshin (n.d.) “Maximum-likelihood estimation of the limiteddependent variable model with endogenous explanatory variable,” http://siteresources.worldbank.org/DEC/Resources/probitiv.pdf Glick, P. , A. Marini and D. Sahn (2005) “Estimating the consequences of changes in fertility on child health and education in Romania: An analysis using twins data,” Cornell Food and Nutrition Policy Program Working Papers 183 Heckman, J. (1993) “ What Has Been Learned About Labor Supply in the Past Twenty Years,” American Economic Review Papers and Proceedings, 83(2), 116-121. Lee, J. (2004) “Sibling Size and Investment in Children’s Education: An Asian Instrument,” IZA Discussion Paper No. 1323. Mroz, T. (1988) “The Sensitivity of an Empirical Model of Married Women’s Hours of Work to Economic and Statistical Assumptions,” Econometrica, 55(4), 765-799. Quian, N. (2004) “Quantity-Quality and the One Child Policy: The Positive Effect of Family Size on School Enrollment in China,” (http://econwww.mit.edu/graduate/candidates/download_res.php?id=130) Rosenzweig, M. and K. Wolpin (1980a) “Testing the Quantity-Quality Fertility Model: The Use of Twins as a Natural Experiment,” Econometrica, 48(1), 227-240 Rosenzweig and Wolpin (1980b) “Life-Cycle Labor Supply and Fertility: A Causal Inferences from Household Models,” Journal of Political Economy, 82(2), 328-348. Rivers, D. and Q. Vuong (1988) “Limited Information Estimators and Exogeneity Tests for Simultaneous Probit Models,” Journal of Econometrics 39, 347-366. Smith, R. and R. Blundell (1986) “An Exogeneity Test for Simultaneous Equation Tobit Model with and Application to Labor Supply,” Econometrica, 54(3), 679-685. Stinner, W. and P. Mader (1975) “Sons, Daughters or Both? An Analysis of Family Sex Composition Preferences in the Philippines,” Demography, 12(1), 67-79. Vere, J. (2005) “Life Cycle Effects of Fertility on Parents’ Labor Supply,” University of Hong Kong.
14 Viitanen, T. (2003) “Children and Their Mothers’ Labor Supply: Evidence for a LargeScale Policy Change in Finland,” Paper presented at the Annual Conference of the European Society for Population Economics, New York, US June 20031 Wongboonsin, K. and V. P. Ruffolo (1995) “Sex Preference for Children in Thailand and Some Other South-East Asian Countries” Asia-Pacific Population Journal, 10(3), 43-62
15 Table 1. Proportion of families that had a third child and average number of children by sex of first two children
Sex of first two children
Proportion that has a third child Mean SD SE
Number of children Mean SD SE
Proportion to sample
(1) One Male, One Female
0.6740
0.4688
0.0042
3.4850
1.5436
0.0315
0.964
(2) Both male
0.7179
0.4500
0.0052
3.6452
1.5994
0.0420
0.432
(3) Both female
0.7180
0.4500
0.0063
3.5575
1.4975
0.0495
0.261
(4) Same Sex
0.7179
0.4500
0.0040
3.6095
1.5592
0.0320
1.037
Difference (4)-(1)
0.0439
0.0058
0.1245
0.0449
Source of basic data: National Statistics Office, Annual Poverty Indicators Survey, 2002
Table 2. Proportion of mothers and fathers working per capita income quintile and number of children, 2002
All typles
Mother Paid work
Father All types
Per capita Income quintile Poorest Lower middle Middle Upper middle Richest
0.534 0.510 0.503 0.555 0.657
0.305 0.334 0.343 0.384 0.384
0.941 0.926 0.901 0.870 0.856
0.550 0.544 0.548 0.550 0.529 0.526 0.519 0.523
0.433 0.433 0.436 0.419 0.411 0.410 0.374 0.398
0.867 0.908 0.918 0.925 0.943 0.948 0.918 0.915
0.5451
0.3489
0.9041
No. of children 2 3 4 5 6 7 8 9 and above Philippines
16
Table 3. Descriptive Statistics Obs Mother working, all types Mother working, paid Father working, all types No of children Age of mother Age of father Education of mother, years Education of father, years Deflated fathers wage earnings, thousand (1994= Non-wage income, thousand (1994=100) Unemployment rate, % domain level Presence of children below 6 years Urban dummy Region 1 dummy Region 2 dummy Region 3 dummy Region 4 dummy Region 5 dummy Region 6 dummy Region 7 dummy Region 8 dummy Region 9 dummy Region 10 dummy Region 11 dummy Region 12 dummy NCR dummy CAR dummy ARMM dummy Caraga dummy
23828 30652 21873 24931 23828 21873 23828 21873 24931 30651 24751 30652 24931 30652 30652 30652 30652 30652 30652 30652 30652 30652 30652 30652 30652 30652 30652 30652 30652
Mean 0.5451 0.3489 0.9041 3.5484 42.9238 45.0971 9.2002 9.0478 1.7588 5.7028 5.4276 0.4961 0.5898 0.0464 0.0376 0.0952 0.1607 0.0533 0.0716 0.0584 0.0533 0.0430 0.0505 0.0509 0.0440 0.1035 0.0441 0.0500 0.0374
Std. Dev. 0.4980 0.4766 0.2944 1.5528 10.8725 10.7714 3.7755 3.7678 4.1080 15.6870 2.6211 0.5000 0.4919 0.2104 0.1902 0.2935 0.3672 0.2246 0.2579 0.2346 0.2247 0.2028 0.2190 0.2197 0.2052 0.3046 0.2053 0.2180 0.1898
Min
Max 0 0 0 2 15 18 0 0 0 0 0.8 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
1 1 1 12 99 99 17 17 313 1578 18 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
17 Table 4. Determinants of labor force participation of mothers, all types of work, 2002 (Two-Step Probit estimates)
Explanatory Variables
Coef.
Predicted No. of children** No. of children Estimated residual of no. of children Age, mother Age, mother squared Years of educ., mother Wage income per capita, father, (000) Non-wage income per capita (000) Unemployment rate, domain level Presence of children below 6 years Urban Region 1 Region 2 Region 3 Region 4 Region 5 Region 6 Region 7 Region 8 Region 9 Region 10 Region 11 Region 12 CAR ARMM Caraga Constant
-0.2543
Psuedo-R2 No of Obs. * Huber-White "Robust" Standard Errors ** Instruments: Both male and both female
0.1934 -0.0020 0.0301 -0.0349 -0.0005 -0.0114 0.1162 -0.1373 0.2819 0.4931 0.0777 0.2900 0.3558 0.4830 0.3715 0.4540 0.0071 0.6521 0.3531 0.5592 0.4719 -0.2662 0.3673 -3.6530 0.0688 23,656
Model 1 Std. Err.* 0.1558
0.0439 0.0004 0.0076 0.0055 0.0022 0.0038 0.2015 0.0192 0.0532 0.0589 0.0392 0.0370 0.0631 0.0493 0.0477 0.0595 0.0520 0.0489 0.0472 0.0623 0.0684 0.0654 0.0526 0.4817
z
Coef.
Model 2 Std. Err.*
z
-1.63
4.40 -4.60 3.97 -6.31 -0.21 -3.03 0.58 -7.15 5.30 8.37 1.98 7.84 5.64 9.79 7.80 7.63 0.14 13.34 7.48 8.98 6.90 -4.07 6.98 -7.58
-0.2536 0.2309 0.1940 -0.0021 0.0302 -0.0353 -0.0005 -0.0115 0.1154 -0.1369 0.2814 0.4927 0.0775 0.2898 0.3554 0.4828 0.3712 0.4539 0.0069 0.6520 0.3528 0.5592 0.4713 -0.2663 0.3671 -3.6647 0.0692 23,656
0.1558 0.1560 0.0439 0.0004 0.0076 0.0056 0.0022 0.0038 0.2016 0.0192 0.0532 0.0589 0.0392 0.0370 0.0631 0.0493 0.0477 0.0595 0.0520 0.0489 0.0472 0.0623 0.0684 0.0654 0.0527 0.4819
-1.63 1.48 4.41 -4.62 3.98 -6.35 -0.23 -3.04 0.57 -7.13 5.29 8.36 1.98 7.84 5.63 9.79 7.79 7.63 0.13 13.33 7.47 8.98 6.89 -4.08 6.97 -7.60
18 Table 5. Determinants of Labor Force Participation of mothers and fathers, 2002 (Full-Information Maximum Likelihood estimates)
Explanatory Variables Labor force participation equation: No. of children Age, mother Age, mother squared Years of educ., mother Wage income per capita, father, (000) Age, father Age, father squared Years of educ., father Non-wage income per capita (000) Unemployment rate, %, domain level Presence of children below 6 years Region 1 Region 2 Region 3 Region 4 Region 5 Region 6 Region 7 Region 8 Region 9 Region 10 Region 11 Region 12 CAR ARMM Caraga Constant
No. of children equation: Both male first two births Both female first two births Age, mother Age, mother squared Years of educ., mother Wage income per capita, father, (000) Age, father Age, father squared Years of educ., father Non-wage income per capita (000) Unemployment rate, domain level Presence of children below 6 years Urban Region 1 Region 2 Region 3 Region 4 Region 5 Region 6 Region 7 Region 8 Region 9 Region 10 Region 11 Region 12 CAR ARMM Caraga Constant LR test of indep. Eqns. (rho=0) Chi2(1) P-value Obs Wald Chi(24)
Coef.
-0.2662 0.1901 -0.0020 0.0271 -0.0340
Mother Std. Err.*
0.4960 0.1076 0.0011 0.0334 0.0085
Coef.
-0.54 1.77 -1.89 0.81 -4.00
-0.0205
0.2235
-0.09
0.0749 -0.0011 -0.0131 -0.0008 -0.0345
0.0457 0.0005 0.0116 0.0030 0.0052
1.64 -2.35 -1.14 -0.27 -6.65
-0.3289 0.1376 0.4158 -0.0242 0.0365 0.3381 0.2665 0.0971 0.4523 0.3895 0.4655 0.3299 0.3063 0.1370 0.3694 0.2286 0.8992
0.0318 0.0770 0.1109 0.0513 0.0490 0.1157 0.0742 0.0697 0.1033 0.0845 0.0781 0.0700 0.0966 0.1013 0.1126 0.0804 0.2582
-10.33 1.79 3.75 -0.47 0.74 2.92 3.59 1.39 4.38 4.61 5.96 4.71 3.17 1.35 3.28 2.84 3.48
0.1365 0.0951
0.0241 0.0273
5.67 3.48
0.2062 -0.0022 -0.0495 -0.0121 -0.0033
0.0062 0.0001 0.0031 0.0020 0.0045
33.44 -33.97 -15.74 -6.01 -0.75
-0.0432 0.1329 -0.2244 0.0131 0.0658 0.4203 0.1897 0.1407 0.3279 0.1155 0.0663 0.0860 0.2865 0.3096 0.3687 0.0258 -0.6012
0.0237 0.0616 0.0597 0.0434 0.0399 0.0582 0.0519 0.0552 0.0612 0.0621 0.0547 0.0537 0.0637 0.0601 0.0608 0.0616 0.1482
-1.82 2.16 -3.76 0.30 1.65 7.22 3.66 2.55 5.36 1.86 1.21 1.60 4.50 5.15 6.07 0.42 -4.06
-0.0008 -0.0108 0.1441 -0.1289 0.2678 0.4603 0.0741 0.2744 0.3415 0.4580 0.3516 0.4330 0.0082 0.6136 0.3338 0.5320 0.4511 -0.2432 0.3440 -3.5276
0.0067 0.0048 0.6920 0.0364 0.0488 0.1928 0.0381 0.0485 0.0755 0.0578 0.0614 0.0481 0.0602 0.1609 0.0652 0.0525 0.0658 0.2204 0.1270 0.7419
-0.12 -2.23 0.21 -3.54 5.49 2.39 1.95 5.66 4.53 7.92 5.73 9.00 0.14 3.81 5.12 10.14 6.85 -1.10 2.71 -4.76
0.1235 0.0628 0.2789 -0.0028 -0.0454 -0.0297
0.0272 0.0432 0.0062 0.0001 0.0034 0.0081
4.54 1.45 44.76 -42.43 -13.31 -3.67
-0.0120 0.0002 1.2841 -0.0055 0.1160 -0.1115 0.0453 0.0699 0.2656 0.1470 0.0934 0.2224 0.0644 0.0185 0.0748 0.2278 0.2846 0.2658 -0.0406 -2.9922
0.0018 0.0040 0.0261 0.0219 0.0558 0.0553 0.0391 0.0362 0.0518 0.0465 0.0487 0.0530 0.0561 0.0497 0.0487 0.0574 0.0548 0.0560 0.0563 0.1514
Father Std. Err.*
z
-6.60 0.06 49.28 -0.25 2.08 -2.02 1.16 1.93 5.13 3.16 1.92 4.20 1.15 0.37 1.53 3.97 5.20 4.75 -0.72 -19.77
0.200 0.657
0.030 0.873
23,656 2,329.09
23,656 1,734.27
z
19 Table 6. Determinants of labor force participation of mothers, by type of work, 2002 (Probit estimates)
Explanatory variables
Coef.
No. of children Age, mother Age, mother squared Years of educ., mother Wage income per capita, father, (000) Non-wage income per capita (000) Unemployment rate, domain level Presence of children below 6 years Urban Region 1 Region 2 Region 3 Region 4 Region 5 Region 6 Region 7 Region 8 Region 9 Region 10 Region 11 Region 12 CAR ARMM Caraga Constant
-0.0231 0.1295 -0.0014 0.0407 -0.0284 0.0023 -0.0116 -0.1812 -0.1354 0.2543 0.5173 0.0673 0.2736 0.2938 0.4486 0.3499 0.4023 -0.0077 0.6472 0.3355 0.5072 0.4052 -0.3269 0.3766 -2.9829
Psuedo-R2 No of Obs. * Huber-White "Robust" Standard Errors
23,656 0.0692
All types of work Std. Err.* z 0.0061 0.0059 0.0001 0.0026 0.0030 0.0011 0.0038 0.0232 0.0192 0.0499 0.0565 0.0386 0.0353 0.0475 0.0437 0.0455 0.0482 0.0511 0.0488 0.0457 0.0514 0.0519 0.0512 0.0522 0.1410
-3.81 21.89 -22.47 15.58 -9.48 2.11 -3.07 -7.82 -7.06 5.10 9.16 1.74 7.75 6.19 10.28 7.69 8.34 -0.15 13.27 7.35 9.86 7.81 -6.38 7.21 -21.15
Mar. Eff. -0.0092 0.0513 -0.0006 0.0161 -0.0113 0.0009 -0.0046 -0.0719 -0.0535 0.0985 0.1919 0.0266 0.1065 0.1133 0.1694 0.1339 0.1526 -0.0031 0.2343 0.1286 0.1886 0.1535 -0.1298 0.1433
Coef. -0.0242 0.1189 -0.0013 0.0603 -0.0069 -0.0031 -0.0055 -0.1444 0.0297 0.1022 0.3211 0.1210 0.2735 0.2370 0.3747 0.3250 0.2982 -0.0386 0.3409 0.2945 0.3290 -0.0103 -0.3867 0.2520 -3.3074 23,656 0.0641
Excluding unpaid Std. Err.* z 0.0061 0.0061 0.0001 0.0026 0.0024 0.0009 0.0037 0.0231 0.0190 0.0498 0.0551 0.0387 0.0353 0.0474 0.0432 0.0455 0.0478 0.0519 0.0472 0.0454 0.0509 0.0516 0.0537 0.0519 0.1451
-4.00 19.46 -19.61 22.83 -2.86 -3.66 -1.46 -6.24 1.56 2.05 5.83 3.13 7.74 5.00 8.68 7.14 6.24 -0.74 7.23 6.49 6.46 -0.20 -7.20 4.86 -22.80
Mar. Eff. -0.0096 0.0469 -0.0005 0.0238 -0.0027 -0.0012 -0.0022 -0.0568 0.0117 0.0405 0.1275 0.0480 0.1086 0.0943 0.1486 0.1291 0.1185 -0.0152 0.1353 0.1171 0.1307 -0.0041 -0.1459 0.1002
20 Table 7. Determinants of labor force participation of mothers, by type of work, 2002 (Probit estimates; with interaction of number of children and per capita income quintile)
Explanatory variables
Coef.
No. of children No. of children x quintile 2 No. of children x quintile 3 No. of children x quintile 4 No. of children x quintile 5 Age, mother Age, mother squared Years of educ., mother Wage income per capita, father, (000) Non-wage income per capita (000) Unemployment rate, domain level Presence of children below 6 years Urban Region 1 Region 2 Region 3 Region 4 Region 5 Region 6 Region 7 Region 8 Region 9 Region 10 Region 11 Region 12 CAR ARMM Caraga Constant
-0.0287 -0.0059 -0.0018 0.0382 0.1393 0.1270 -0.0014 0.0296 -0.0396 -0.0029 -0.0132 -0.1654 -0.1536 0.3128 0.5667 0.1080 0.3061 0.3615 0.5122 0.3986 0.4614 0.0411 0.7152 0.4088 0.5635 0.4290 -0.3068 0.4500 -2.8182
Psuedo-R2 No of Obs.
All types of work Std. Err.* z 0.0067 0.0059 0.0068 0.0079 0.0123 0.0060 0.0001 0.0028 0.0036 0.0010 0.0038 0.0234 0.0196 0.0503 0.0568 0.0390 0.0357 0.0481 0.0443 0.0460 0.0487 0.0513 0.0495 0.0462 0.0519 0.0525 0.0514 0.0528 0.1422
Mar. Eff.
-4.27 -1.01 -0.26 4.85 11.36 21.30 -22.24 10.60 -11.10 -2.79 -3.49 -7.08 -7.85 6.22 9.98 2.77 8.57 7.52 11.55 8.66 9.47 0.80 14.46 8.84 10.86 8.17 -5.97 8.53 -19.81
-0.0114 -0.0023 -0.0007 0.0152 0.0552 0.0503 -0.0006 0.0117 -0.0157 -0.0011 -0.0052 -0.0656 -0.0607 0.1202 0.2080 0.0425 0.1188 0.1380 0.1912 0.1514 0.1732 0.0162 0.2548 0.1549 0.2071 0.1618 -0.1219 0.1690
Excluding unpaid Std. Err.* z
Coef. -0.0509 0.0259 0.0372 0.0821 0.1825 0.1156 -0.0013 0.0447 -0.0200 -0.0097 -0.0074 -0.1095 -0.0123 0.1691 0.3766 0.1531 0.3083 0.3381 0.4562 0.3908 0.3886 0.0317 0.4418 0.3871 0.4121 0.0181 -0.3704 0.3604 -3.0899
0.0753 23,656
0.0068 0.0060 0.0069 0.0079 0.0123 0.0062 0.0001 0.0028 0.0030 0.0013 0.0038 0.0234 0.0195 0.0503 0.0555 0.0391 0.0357 0.0481 0.0439 0.0461 0.0485 0.0522 0.0481 0.0461 0.0517 0.0519 0.0537 0.0524 0.1464
-7.46 4.32 5.43 10.45 14.82 18.74 -19.35 15.77 -6.68 -7.41 -1.97 -4.69 -0.63 3.36 6.78 3.92 8.63 7.03 10.40 8.48 8.01 0.61 9.18 8.40 7.98 0.35 -6.90 6.87 -21.11
0.0726 23,656
* Huber-White "Robust" Standard Errors
Table 8. Impact on labor force participation (LFP) of mothers and fathers by per capita income quintile as % of recorded LFP Mother All types
Paid
Father All types
Average
-1.68
-2.13
0.00 *
Poorest Lower middle Middle Upper middle Richest
-2.12 -2.12 * -2.12 * 0.69 6.68
-5.68 -2.43 -1.26 2.45 8.52
0.00 * 0.33 0.60 0.43 1.16
* insignificant, assumed same as base case Computed from Tables 7 and 9
Mar. Eff. -0.0201 0.0102 0.0147 0.0324 0.0720 0.0456 -0.0005 0.0176 -0.0079 -0.0038 -0.0029 -0.0431 -0.0048 0.0672 0.1493 0.0608 0.1224 0.1342 0.1802 0.1549 0.1540 0.0125 0.1746 0.1534 0.1631 0.0072 -0.1401 0.1430
21 Table 9. Determinants of labor force participation of fathers, 2002 (Two-Step Probit estimates)
Explanatory Variables
Coef.
Predicted No. of children** No. of children Estimated residual of no. of children Age, father Age, father squared Years of educ., father Nonwage income per capita, (000) Unemployment rate, domain level Urban Region 1 Region 2 Region 3 Region 4 Region 5 Region 6 Region 7 Region 8 Region 9 Region 10 Region 11 Region 12 CAR ARMM Caraga Constant
-0.0228
Psuedo-R2 No of Obs. * Huber-White "Robust" Standard Errors ** Instruments: Both male and both female
0.0749 -0.0011 -0.0132 -0.0008 -0.0346 -0.3291 0.1385 0.4160 -0.0242 0.0361 0.3368 0.2670 0.0979 0.4519 0.3893 0.4659 0.3312 0.3070 0.1369 0.3719 0.2288 0.9118 0.1619 21,709
Model 1 Std. Err.* 0.2092
0.0441 0.0005 0.0111 0.0028 0.0052 0.0329 0.0773 0.1058 0.0514 0.0490 0.1157 0.0753 0.0695 0.1054 0.0859 0.0790 0.0711 0.0973 0.1000 0.1135 0.0805 0.2473
z
Coef.
Model 2 Std. Err.*
z
-0.11
1.70 -2.41 -1.19 -0.28 -6.61 -10.01 1.79 3.93 -0.47 0.74 2.91 3.55 1.41 4.29 4.53 5.90 4.66 3.16 1.37 3.28 2.84 3.69
-0.0189 0.0344 0.0747 -0.0011 -0.0131 -0.0008 -0.0345 -0.3293 0.1376 0.4168 -0.0242 0.0364 0.3379 0.2665 0.0970 0.4524 0.3898 0.4661 0.3302 0.3063 0.1367 0.3693 0.2289 0.9015 0.1621 21,709
0.2092 0.2096 0.0441 0.0005 0.0111 0.0028 0.0052 0.0329 0.0773 0.1058 0.0513 0.0490 0.1158 0.0753 0.0695 0.1056 0.0859 0.0789 0.0711 0.0973 0.1000 0.1134 0.0807 0.2459
-0.09 0.16 1.69 -2.40 -1.18 -0.27 -6.61 -10.01 1.78 3.94 -0.47 0.74 2.92 3.54 1.40 4.29 4.54 5.90 4.64 3.15 1.37 3.26 2.84 3.67
22 Table 10. Determinants of labor force participation of fathers, 2002 (Probit estimates)
Explanatory variables
Without per capita income quintile Coef. Std. Err.* z Mar. Eff.
No. of children No. of children x quintile 2 No. of children x quintile 3 No. of children x quintile 4 No. of children x quintile 5 Age, father Age, father squared Years of educ., father Nonwage income per capita, (000) Unemployment rate, domain level Urban Region 1 Region 2 Region 3 Region 4 Region 5 Region 6 Region 7 Region 8 Region 9 Region 10 Region 11 Region 12 CAR ARMM Caraga Constant Psuedo-R2 No of Obs. * Huber-White "Robust" Standard Errors
0.0155
0.0091
1.72
0.0021
0.0676 -0.0011 -0.0114 -0.0004 -0.0344 -0.3278 0.1330 0.4244 -0.0246 0.0342 0.3233 0.2600 0.0922 0.4411 0.3858 0.4636 0.3272 0.2966 0.1261 0.3567 0.2279 0.9203
0.0088 0.0001 0.0039 0.0012 0.0052 0.0317 0.0721 0.0950 0.0513 0.0473 0.0752 0.0642 0.0633 0.0804 0.0827 0.0780 0.0690 0.0783 0.0751 0.0856 0.0806 0.2175
7.71 -12.08 -2.92 -0.30 -6.64 -10.35 1.85 4.47 -0.48 0.72 4.30 4.05 1.46 5.48 4.67 5.95 4.75 3.79 1.68 4.17 2.83 4.23
0.0091 -0.0001 -0.0015 0.0000 -0.0046 -0.0426 0.0163 0.0425 -0.0034 0.0045 0.0350 0.0296 0.0117 0.0440 0.0399 0.0456 0.0354 0.0325 0.0156 0.0380 0.0262
0.1621 21,709
With per capita income quintile Coef. Std. Err.* z Mar. Eff. -0.0006 0.0229 0.0400 0.0280 0.0741 0.0644 -0.0010 -0.0183 -0.0016 -0.0351 -0.3511 0.1628 0.4415 -0.0128 0.0460 0.3700 0.2885 0.1187 0.4781 0.4228 0.5100 0.3663 0.3368 0.1385 0.3772 0.2794 1.0482 0.1640 21,709
0.0102 0.0100 0.0110 0.0119 0.0161 0.0088 0.0001 0.0044 0.0012 0.0052 0.0325 0.0726 0.0954 0.0513 0.0475 0.0768 0.0647 0.0638 0.0817 0.0834 0.0793 0.0696 0.0788 0.0753 0.0859 0.0819 0.2216
-0.06 2.30 3.65 2.35 4.61 7.30 -11.82 -4.19 -1.38 -6.75 -10.81 2.24 4.63 -0.25 0.97 4.82 4.46 1.86 5.85 5.07 6.43 5.26 4.28 1.84 4.39 3.41 4.73
-0.0001 0.0031 0.0054 0.0038 0.0099 0.0086 -0.0001 -0.0025 -0.0002 -0.0047 -0.0454 0.0195 0.0435 -0.0017 0.0060 0.0387 0.0321 0.0147 0.0463 0.0424 0.0485 0.0384 0.0358 0.0169 0.0395 0.0309
23 Table 11. Determinants of wage income of mothers (2-Stage Tobit estimates) Model 1 Std. Err.*
Explanatory Variables
Coef.
No. of children Estimated residual of no. of children** Predicted no. of children** Age, mother Age, mother squared Years of educ., mother Urban Region 1 Region 2 Region 3 Region 4 Region 5 Region 6 Region 7 Region 8 Region 9 Region 10 Region 11 Region 12 CAR ARMM Caraga Constant
-0.1669 0.1168
0.1841 0.1842
-0.91 0.63
0.1040 -0.0010 0.2137 0.3981 -0.6006 -0.6867 -0.3019 -0.2733 -0.6271 -0.5662 -0.3616 -0.5078 -0.3843 -0.7846 -0.5659 -0.5283 -0.3068 -0.0376 -0.7451 -1.7328
0.0341 0.0004 0.0119 0.0311 0.0817 0.0745 0.0581 0.0543 0.1074 0.0751 0.0748 0.0993 0.0918 0.0700 0.0736 0.0890 0.0960 0.1202 0.0821 0.2574
3.05 -2.58 18.02 12.81 -7.35 -9.22 -5.20 -5.03 -5.84 -7.54 -4.84 -5.12 -4.18 -11.21 -7.69 -5.93 -3.20 -0.31 -9.07 -6.73
0.9259
0.0092
Censoring parameter Observations * Huber-White "Robust" Standard Errors ** Instruments: Both male and both female
5,540
t
Coef.
Model 2 Std. Err.*
-0.1536 0.0967 -0.0009 0.2151 0.3999 -0.6039 -0.6842 -0.3105 -0.2755 -0.6307 -0.5684 -0.3611 -0.5111 -0.3871 -0.7918 -0.5669 -0.5368 -0.3167 -0.0969 -0.7426 -1.6403
0.1846 0.0342 0.0004 0.0119 0.0312 0.0819 0.0747 0.0582 0.0545 0.1078 0.0753 0.0750 0.0996 0.0921 0.0702 0.0738 0.0893 0.0962 0.1202 0.0824 0.2575
0.9287
0.0093
5,540
t
-0.83 2.83 -2.36 18.09 12.82 -7.37 -9.15 -5.33 -5.06 -5.85 -7.55 -4.82 -5.13 -4.20 -11.29 -7.68 -6.01 -3.29 -0.81 -9.01 -6.37
24 Table 12. Determinants of wage income of mothers (Tobit and OLS estimates) Tobit Estimates Without per capita income quintile Without per capita income quintile Coef. Std. Err. t Mar. Eff.* Coef. Std. Err. t Mar. Eff.*
Coef.
-0.0504
0.0088
-5.75
-0.0500
-0.0510
0.0094
-5.45
0.0835 -0.0008 0.2208 0.4059 -0.6217 -0.6688 -0.3067 -0.2857 -0.6793 -0.5945 -0.3814 -0.5520 -0.4009 -0.7971 -0.5804 -0.5609 -0.3457 -0.0724 -0.7548 -1.7960
0.0111 0.0001 0.0035 0.0286 0.0746 0.0690 0.0576 0.0507 0.0690 0.0603 0.0679 0.0707 0.0880 0.0671 0.0700 0.0727 0.0738 0.1070 0.0807 0.2373
7.55 -6.32 62.78 14.20 -8.34 -9.70 -5.33 -5.64 -9.84 -9.86 -5.61 -7.80 -4.55 -11.87 -8.29 -7.71 -4.68 -0.68 -9.35 -7.57
0.0829 -0.0008 0.2192 0.4022 -0.6104 -0.6561 -0.3034 -0.2829 -0.6661 -0.5849 -0.3766 -0.5431 -0.3956 -0.7789 -0.5706 -0.5516 -0.3415 -0.0718 -0.7378 -1.7828
0.0860 -0.0008 0.2214 0.4054 -0.6213 -0.6905 -0.3102 -0.3021 -0.6849 -0.6067 -0.4009 -0.5898 -0.4057 -0.7922 -0.5909 -0.5539 -0.3463 -0.0706 -0.7887 -1.8615
0.0120 0.0001 0.0038 0.0313 0.0758 0.0682 0.0531 0.0454 0.0646 0.0569 0.0682 0.0724 0.0812 0.0660 0.0695 0.0667 0.0728 0.0867 0.0898 0.2566
7.18 -6.03 58.91 12.96 -8.20 -10.12 -5.84 -6.66 -10.59 -10.66 -5.88 -8.15 -5.00 -12.00 -8.50 -8.30 -4.76 -0.81 -8.78 -7.26
Censoring parameter
0.9260
0.0092
Pseudo R2/R-square Observations
0.2250 5,540
Explanatory variables No. of children No. of children x quintile 2 No. of children x quintile 3 No. of children x quintile 4 No. of children x quintile 5 Age, mother Age, mother squared Years of educ., mother Urban Region 1 Region 2 Region 3 Region 4 Region 5 Region 6 Region 7 Region 8 Region 9 Region 10 Region 11 Region 12 CAR ARMM Caraga Constant
-0.1277 0.0599 0.1485 0.2822 0.4613 0.0569 0.2603 -0.0006 0.1317 -0.3576 -0.4331 -0.1490 -0.1422 -0.3030 -0.3031 -0.1783 -0.2344 -0.1046 -0.4174 -0.2384 -0.2276 -0.2538 -0.1131 -0.3702 -0.5335
0.0089 0.0087 0.0099 0.0104 0.0132 0.0098 0.0257 0.0001 0.0039 0.0666 0.0615 0.0515 0.0452 0.0621 0.0541 0.0606 0.0634 0.0786 0.0604 0.0628 0.0652 0.0656 0.0954 0.0724 0.2131
0.8219
0.0082
-14.37 6.89 15.07 27.01 35.06 5.79 10.11 -5.66 33.54 -5.37 -7.04 -2.89 -3.15 -4.88 -5.60 -2.94 -3.70 -1.33 -6.91 -3.80 -3.49 -3.87 -1.19 -5.12 -2.50
0.2893 5,540
-0.1273 0.0598 0.1481 0.2814 0.4599 0.0567 0.2594 -0.0006 0.1313 -0.3556 -0.4304 -0.1484 -0.1417 -0.3015 -0.3016 -0.1776 -0.2333 -0.1043 -0.4148 -0.2374 -0.2266 -0.2527 -0.1127 -0.3680 -0.5319
OLS (Robust SE) Std. Err.
0.5224 5,540
* Unconditional expected value
Table 13. Impact on wage income of mothers and fathers by per capita income quintile Mothers As % of inc. Abs. value Average Poorest Lower middle Middle Upper middle Richest
Fathers As % of inc. Abs. value
-5.0
-1,010
1.1
233
-12.7 -6.8 2.1 15.4 33.3
-659 -598 360 6,200 25,736
-6.0 5.1 12.5 18.7 35.4
-76 93 394 1,762 12,538
* insignificant, assumed same as base case Computed from Tables 12 and 15
t
25 Table 14. Determinants of wage income of fathers (2-Stage Tobit estimates) Model 1 Std. Err.*
Explanatory Variables
Coef.
No. of children Estimated residual of no. of children** Predicted no. of children** Age, father Age, father squared Years of educ., father Urban Region 1 Region 2 Region 3 Region 4 Region 5 Region 6 Region 7 Region 8 Region 9 Region 10 Region 11 Region 12 CAR ARMM Caraga Constant
-0.0852 0.0959
0.1149 0.1149
-0.74 0.83
0.0433 -0.0005 0.1318 0.4073 -0.5923 -0.6026 -0.1668 -0.2434 -0.6554 -0.4898 -0.3053 -0.4787 -0.2784 -0.6832 -0.4213 -0.4156 -0.1581 -0.2271 -0.6279 0.7873
0.0246 0.0003 0.0067 0.0184 0.0475 0.0502 0.0321 0.0306 0.0661 0.0436 0.0437 0.0587 0.0483 0.0407 0.0415 0.0553 0.0580 0.0770 0.0466 0.1506
1.76 -1.75 19.53 22.12 -12.48 -12.02 -5.19 -7.96 -9.92 -11.25 -6.99 -8.15 -5.76 -16.78 -10.14 -7.51 -2.73 -2.95 -13.48 5.23
Censoring parameter
0.7830
0.0054
Observations
21,873
* Huber-White "Robust" Standard Errors ** Instruments: Both male and both female
t
Coef.
Model 2 Std. Err.*
-0.3266 0.0707 -0.0008 0.1240 0.3845 -0.5495 -0.6397 -0.1510 -0.2206 -0.5501 -0.4193 -0.2641 -0.3933 -0.2270 -0.6611 -0.3839 -0.3610 -0.0951 -0.1457 -0.6055 1.1147
0.0335 0.0076 0.0001 0.0026 0.0177 0.0451 0.0455 0.0323 0.0294 0.0425 0.0374 0.0394 0.0438 0.0461 0.0395 0.0405 0.0447 0.0464 0.0637 0.0464 0.1397
0.7779
0.0054
21,873
t
-9.75 9.34 -9.10 47.21 21.78 -12.19 -14.07 -4.67 -7.50 -12.95 -11.22 -6.70 -8.98 -4.93 -16.72 -9.47 -8.07 -2.05 -2.29 -13.05 7.98
26 Table 15. Determinants of wage income of fathers (Tobit and OLS estimates)
Explanatory variables No. of children No. of children x quintile 2 No. of children x quintile 3 No. of children x quintile 4 No. of children x quintile 5 Age, father Age, father squared Years of educ., father Urban Region 1 Region 2 Region 3 Region 4 Region 5 Region 6 Region 7 Region 8 Region 9 Region 10 Region 11 Region 12 CAR ARMM Caraga Constant
Tobit Estimates Without per capita income quintile Without per capita income quintile Coef. Std. Err. t Mar. Eff.* Coef. Std. Err. t Mar. Eff.* 0.0107
0.0050
2.13
0.0107
0.0234 -0.0002 0.1371 0.4126 -0.6071 -0.5831 -0.1700 -0.2517 -0.6997 -0.5101 -0.3225 -0.5131 -0.2930 -0.6931 -0.4313 -0.4444 -0.1889 -0.2648 -0.6339 0.8426
0.0062 0.0001 0.0023 0.0172 0.0440 0.0444 0.0319 0.0289 0.0394 0.0361 0.0386 0.0417 0.0451 0.0389 0.0398 0.0433 0.0447 0.0623 0.0460 0.1353
3.75 -3.59 59.90 23.93 -13.79 -13.14 -5.33 -8.72 -17.77 -14.13 -8.37 -12.30 -6.50 -17.80 -10.84 -10.26 -4.22 -4.25 -13.77 6.23
0.0234 -0.0002 0.1371 0.4124 -0.6061 -0.5822 -0.1699 -0.2516 -0.6984 -0.5095 -0.3222 -0.5125 -0.2928 -0.6918 -0.4308 -0.4439 -0.1888 -0.2646 -0.6328 0.8422
Censoring parameter
0.7830
0.0054
Pseudo R2/R-square Observations
0.1801 10,995
* Unconditional expected value
-0.0598 0.1104 0.1844 0.2472 0.4137 0.0056 -0.0002 0.0777 0.2640 -0.4174 -0.4436 -0.0957 -0.1706 -0.3781 -0.2932 -0.1581 -0.3042 -0.0702 -0.4061 -0.2128 -0.2041 -0.1622 -0.2480 -0.3397 1.6545
0.0049 0.0047 0.0053 0.0061 0.0082 0.0055 0.0001 0.0023 0.0156 0.0393 0.0395 0.0284 0.0257 0.0355 0.0324 0.0344 0.0373 0.0403 0.0350 0.0356 0.0387 0.0397 0.0553 0.0413 0.1211
0.6948
0.0048
0.2612 10,995
-12.10 23.31 34.86 40.26 50.74 1.01 -2.85 33.35 16.94 -10.63 -11.22 -3.37 -6.64 -10.64 -9.06 -4.60 -8.15 -1.74 -11.60 -5.98 -5.27 -4.09 -4.49 -8.23 13.67
-0.0598 0.1104 0.1844 0.2472 0.4136 0.0056 -0.0002 0.0777 0.2640 -0.4173 -0.4435 -0.0957 -0.1706 -0.3780 -0.2932 -0.1581 -0.3041 -0.0702 -0.4060 -0.2128 -0.2041 -0.1622 -0.2479 -0.3397 1.6544
OLS (Robust SE) Coef.
Std. Err.
t
0.0113
0.0054
2.09
0.0228 -0.0002 0.1381 0.4196 -0.6117 -0.5902 -0.1680 -0.2531 -0.7075 -0.5130 -0.3220 -0.5290 -0.2871 -0.6992 -0.4305 -0.4438 -0.1914 -0.2682 -0.6347 0.8335
0.0071 0.0001 0.0025 0.0189 0.0460 0.0493 0.0260 0.0246 0.0407 0.0343 0.0327 0.0435 0.0387 0.0413 0.0370 0.0414 0.0460 0.0624 0.0473 0.1494
3.21 -3.00 55.06 22.24 -13.30 -11.97 -6.46 -10.30 -17.38 -14.95 -9.85 -12.15 -7.42 -16.93 -11.63 -10.73 -4.16 -4.30 -13.41 5.58
0.3986 10,995